Explicit Measures Podcast

Mike & Tommy tackle whether AI agents can solve what the Power BI admin portal can't—giving you a true tenant-level, workspace-free view of who can access what in Fabric. They break down Taylor B.'s real-world agent that stitches together audit logs, model APIs, grants, direct links, and app audiences into a central governance store, and weigh in on whether Microsoft's native governance story is ready or still a DIY puzzle.

From Purview's role to the risks of letting agents reason over permissions, they explore what it would take to make data governance in Fabric actually feel complete—and what builders like Taylor should ship now versus wait for.

Mike & Tommy tackle the often-overlooked "outer context" problem in agentic Fabric development — exploring why request tracking systems like Notion, GitHub Issues, or even ADO matter more than most teams realize, and how building a proper project harness could be the difference between an agent that delivers and one that drifts.

They break down how to map these workflows across corporate tool stacks — Loop, OneNote, SharePoint — and what a minimum viable harness actually looks like in practice.

Mike & Tommy tackle the surprisingly tricky problem of tracking Power BI app usage at scale, exploring why app-level telemetry is harder to surface than report-level metrics and how teams can map usage back to districts using Entra ID, Admin APIs, and audit logs.

They break down which telemetry sources are actually viable, how to avoid common pitfalls like audience filters hiding true reach and shared devices skewing counts, and lay out a scalable architecture for ~9,000 users across 70 districts built around a centralized semantic model with incremental refresh.

Mike & Tommy tackle why organizations are still dragging their feet on AI adoption even with Microsoft Fabric sitting right in front of them — weighing in on whether the real blockers are skills, data quality, governance, or just unclear ROI.

They break down what "AI-ready" actually looks like in a Fabric and Power BI environment, how to tell AI strategy from AI theater, and what a realistic 30/60/90-day plan looks like to move from experimentation to production.

Get in touch:

Mike & Tommy dive into the second half of the AI self-service analytics story, tackling governance, skills, and validation — and why giving everyone the ability to ask questions of data raises the stakes, not just the speed.

They break down what a "Skill" actually means in Power BI terms, who owns the wrong answer when Claude writes the query, and how BI teams can build validation checkpoints that create trust without grinding the business to a halt.

Mike & Tommy dive into self-service analytics with AI, exploring how tools like Claude are changing the way business users interact with data — and whether speed and access come at the cost of trust and governance.

They break down what semantic models need to look like when chat is the primary interface, who owns accountability when AI gives a wrong answer, and what practical guardrails BI teams should put in place before scaling AI-assisted analytics.

Mike & Tommy tackle the growing tension between individual and team-owned Microsoft Fabric Agent Skills, exploring how BI teams should organize, govern, and promote skills without killing innovation or creating shadow BI.

They break down the full skill lifecycle — from personal experiments to certified team assets — and land on practical governance guardrails that are lightweight enough to actually stick.

Get in touch:

Mike & Tommy tackle whether Power BI developers are quietly becoming professional QA testers in the age of Microsoft MCP, weighing in on what separates a true senior developer from a junior when AI writes the first draft.

They explore how the developer role is shifting, who owns accountability when AI-generated measures ship broken, and what skills still matter when the tool can do the typing.

Mike & Tommy dive into RayFin, Microsoft's new AI-first product for building, deploying, and governing agents inside Microsoft Fabric, exploring whether this is a true shift in how BI teams work or just another layer of hype on day one.

They break down what RayFin actually changes for semantic models, reports, and pipelines, tackle the real governance risks when AI can build and deploy at platform speed, and land on the practical first steps a Fabric team should take before turning it loose in their tenant.

Mike & Tommy dive into Microsoft Build 2026, breaking down what the "agentic apps" announcement actually means for Power BI and Fabric teams—and whether this is a real architecture shift or just rebranded copilots.

They weigh in on how Fabric Data Factory orchestration, Microsoft Databases, and Agent Skills change the day-to-day for BI developers and data engineers, and what governance guardrails teams need before anything touches production data.

Mike & Tommy dive into CI/CD automation with agents in Microsoft Fabric, exploring how agentic workflows are reshaping deployment pipelines, whether AI-driven deployments introduce more speed or more risk, and what guardrails teams need before letting agents touch production workspaces.

Mike & Tommy dive into Claude Design meets Power BI Embedded, exploring whether AI-generated UX is a shortcut or a quality risk, how semantic models stay the source of truth when LLMs scaffold embedded apps, and what guardrails belong on every AI-assisted analytics project.

Mike & Tommy tackle Microsoft Fabric hints to work faster, exploring whether teams are wasting time due to self-inflicted bad workflows, poor reuse, and confusing busy work with shipping value. They break down which repeatable tasks should be standardized first, when "just build it" becomes technical debt, and what practical Fabric habits listeners can adopt this week to stop the waste and start delivering faster.

Get in touch:

Mike & Tommy explore agentic skill and report design, diving into the "Five Minutes to Wow" framework and whether AI agents accelerate great design or amplify bad habits. They confront questions about who owns design decisions when the AI makes the call, how to balance speed with craftsmanship, and what guardrails prevent self-service clutter—ultimately wrapping up with practical takeaways for anyone opening Power BI Desktop tomorrow.

Mike & Tommy dive into whether agentic tooling should be driving your Knowledge Center, exploring if docs, support, and governance content can—or should—be fully automated by AI, and what that means for the future relevance of your CoE.

Get in touch:

Mike & Tommy explore the importance of agentic skills for the Fabric developer, weighing whether Skills are a productivity breakthrough or another governance risk, and discussing how to build, version, and share Skills without sacrificing engineering discipline.

Mike & Tommy dive into Databricks Genie and the growing hype around data agents, exploring whether the real challenge is natural language chat or the semantic layer underneath—and what Power BI teams must fix before any AI agent can deliver trusted, governed answers at scale.

Mike & Tommy weigh in on whether bookmarks in Power BI are still worth the maintenance burden, exploring how field parameters, slicers, and modern features have changed the game, and when you should still reach for bookmarks versus simplifying your data model instead.

Mike & Tommy dive into Building Agentic Tools for Microsoft Fabric With Alex Powers, exploring Microsoft's Alex Powers' new tool 'Taskflow Assistant' for Microsoft Fabric and how AI is transforming the way professionals work with Fabric workflows, discussing whether we're helping users master the platform or avoid learning it altogether.

Mike and Tommy review the most underrated Fabric features based on buzz and client usage

Get in touch:

Mike & Tommy tackle the overfed golden dataset problem—when a single "source of truth" becomes a 100-table, 400-measure monster that takes four minutes to open. They explore why teams fall into the trap of cramming everything into one model, how fear of discrepant numbers drives over-centralization, and what it takes to split a bloated semantic model without breaking executive trust or reverting to data chaos.

Get in touch:

This week on Explicit Measures, we tackle a critical "last mile" problem in Business Intelligence: how do you add narrative context to Power BI reports when the storytellers aren't Power BI users? Our listener has built monthly snapshot reports delivered to multiple business units, but marketing leads need to annotate these reports with context before they reach their final audience. We explore the gap between data delivery and storytelling, challenge assumptions about tool selection, and discuss practical patterns for integrating qualitative commentary with quantitative metrics—without breaking your data governance or forcing non-technical users into unfamiliar platforms.

Get in touch:

Mike & Tommy weigh in on whether Databricks and Microsoft Fabric are converging into direct competitors, exploring how Databricks' push into BI with Genie and AI capabilities is closing the gap on Power BI's presentation layer. They question whether "end-to-end" platforms are the future or just feature bloat, discuss where semantic models should live in a modern data stack, and help teams decide when to bet on one platform versus embracing the dual-stack reality.

Get in touch:

Mike & Tommy dive into improving AI skills for Fabric developers, exploring whether to start with prompt engineering or semantic modeling fundamentals, how to build real agents without organizational buy-in, and when AI tools actually add value beyond well-built reports. They tackle the tension between learning cutting-edge capabilities and mastering core BI craft, offering a practical 90-day roadmap for developers stuck in slow-adoption environments.

Get in touch:

Mike & Tommy dive into Joachim's repo-based golden semantic model architecture, exploring whether AI-assisted development is making the classic thin-report pattern obsolete and how upcoming MCP servers for report modeling might reshape architectural decisions. They weigh the trade-offs between file-based AI edits and application-layer safety, discuss when splitting reports makes sense, and tackle the question of whether Power BI developers are becoming software engineers by another name.

Get in touch:

Mike & Tommy recap their experiences at FabCon 2026, sharing key takeaways from sessions, conversations with the community, and highlights from the second day of the conference. They discuss standout moments, emerging trends in the Fabric ecosystem, and what these developments mean for practitioners working with Power BI and the broader Microsoft data platform.

Get in touch:

We are covering ALL of Fabric Con 2026 and highlighting the updates.

Context:

Mike & Tommy dive into Power BI subscriptions as proactive workflow tools, exploring whether automated data nudges can outperform interactive dashboards for busy professionals. Using a real-world data quality monitoring example, they question when the best BI solution is the one users never open, examining the balance between "push" alerts and "pull" exploration, and how subscription-driven reports can save hours while maintaining data hygiene.

Get in touch:

Mike & Tommy dive into DAX complexity and Power Query ETL tips, tackling how to structure complex measures that balance readability and performance across nested time intelligence, dynamic filters, and multi-grain data. They explore advanced Power Query techniques for reducing refresh times in Fabric Dataflows Gen2 and SharePoint integrations, with practical advice on optimizing step folding and reusing logic across queries.

Get in touch:

Mike & Tommy tackle the messy reality of CI/CD for Power BI in the age of TMDL and PBIR, weighing whether Fabric Deployment Pipelines or Azure DevOps—or both—offer the right path forward. They explore why selective file commits feel impossible despite text-based artifacts, when each approach makes sense, and whether mixing deployment methods creates more problems than it solves. The conversation cuts through DevOps hype to reveal what actually works for BI teams navigating source control, pull requests, and environment promotion.

Get in touch:

Mike & Tommy tackle Git best practices for Power BI teams, exploring how to reduce diff noise from slicer selections and matrix expansions, whether renaming visual folders is brilliant or risky, and how small teams can maintain clean, readable source control without losing their minds. They weigh practical tactics like commit checklists, folder naming conventions, and the trade-offs between tracking every click versus only intentional changes.

Get in touch:

Mike & Tommy tackle the critical security distinctions between Publish to Web and Power BI Embedded when exposing reports publicly, questioning whether Embedded actually provides meaningful protection over Publish to Web when authentication isn't required. They explore how Row-Level Security behaves differently in each approach, whether URL filters can be locked down, and what organizations should actually do when building public-facing dashboards—including when to aggregate data upstream, disable granular access entirely, or reconsider whether the data should be public at all.Get in touch:

Description:

In this mailbag episode, we tackle a question from Jacob, a BI Manager and father of three who's been building his Power BI consulting side hustle since late 2023. With a full-time role, a stay-at-home spouse, and limited time to market his services, Jacob asks: How do you scale a consulting business while balancing family responsibilities? We discuss marketing strategies for time-strapped consultants, targeting small to mid-sized businesses, pricing and positioning, and the tension between growth and sustainability. Whether you're considering a side hustle or already running one, this episode offers practical advice on building a consulting practice that fits your life.

Get in touch:

Mike & Tommy dive into what comes after a successful Power BI rollout, exploring whether the next phase is about building more reports or maturing the data platform beneath them. They question if 200 users and 300 KPIs signal maturity or hidden sprawl, discuss the shift from tool deployment to organizational capability, and outline a practical 3-year roadmap focused on governance, scalability, and turning BI into the company's decision engine.

Get in touch:

Mike & Tommy tackle how to submit Fabric ideas that actually get noticed, exploring what separates impactful feedback from noise, why most suggestions get ignored, and how to read Microsoft's roadmap for emerging opportunities. They discuss new data types, future-proofing data engineering beyond AI hype, and practical strategies for becoming a better feedback citizen in the Fabric ecosystem.

Get in touch:

Mike & Tommy tackle the challenge of kicking off Fabric with a clean slate, exploring how to prevent model fragmentation and build a culture of shared semantic models from day one. They discuss whether to rebuild from scratch or incrementally consolidate, why teams resist reusability, and how to establish governance that encourages collaboration without becoming a bottleneck. The episode delivers practical strategies for defining core business entities, creating a semantic model catalog, and building the organizational discipline needed to maintain a true single source of truth.

Starting out with Power BI, our company's reporting team has been using import mode exclusively and we have not been reusing models very commonly. There is a lot of redundant data storage and I'm sure there are many conflicting DAX definitions and model relationships. I've started learning about Fabric and I'm realizing we will need to rebuild some semantic models from the ground up.

Thing is, I'm not sure our team will have the patience to do things right. I suspect they will dive in and start creating more redundancies and our work will continue to be fragmented. How can I kick things off on the right foot and influence the team culture to work together with a single source of truth rather than individual sources?

Get in touch:

Mike & Tommy tackle AI-assisted TMDL workflows and the hot reload problem, exploring whether direct file editing with AI tools like Copilot is the future of Power BI development or a recipe for broken models. They weigh the tension between "move logic upstream" best practices and the brutal close-reopen cycle when TMDL changes introduce errors, debating whether Tabular Editor 3, MCP servers, or a native Microsoft solution offers the best path forward for safe, validated bulk refactoring.Get in touch:

Are we only Direct Lake now? With Microsoft Fabric's ability to use Direct Lake with Semantic Models, where is the need for Import in Power BI, or rather where do we push people?

Get in touch:

In this mailbag episode, we tackle a common challenge in Power BI report design: filter overload. When 15-20 slicers consume a third of your report page, are you empowering users or overwhelming them? We discuss practical guidelines for slicer design, the difference between "set once" and "frequently changed" filters, strategies for moving your team toward more mature reporting practices, and whether there's ever a justifiable reason to exceed 2-3 slicers per page.

Get in touch:

In this episode, we tackle a critical gap between Power BI's Field Parameters and Excel's live connection capabilities. Inspired by a mailbag submission from Eivind Haugen, we explore the tension between building semantic models that look great in Power BI reports versus models that serve all consumers—including the often-overlooked Excel power users.

We dig into why Field Parameters break the Excel experience, when Calculation Groups might be the better choice, and how to design semantic models that truly serve as a single source of truth across multiple consumption tools. If you're building models that connect to Excel, this conversation will change how you think about feature selection and governance.

Get in touch:

In this episode, Mike and Tommy explore whether the traditional "Report Developer" role still exists in the age of AI and Microsoft Fabric. They discuss what skills organizations should actually be hiring for, how the role has evolved from pixel-perfect report building to semantic modeling and data architecture, and whether AI tools like Copilot are changing what we expect from BI professionals. The conversation challenges conventional hiring practices and offers practical guidance on future-proofing BI teams.

Get in touch:

Mike & Tommy weigh in on whether Microsoft Fabric has earned the same trust as Power BI, exploring if it's truly dependable for small businesses and enterprises alike, and examining what trust even means for an all-in-one platform that's still rapidly evolving.

Mike & Tommy explore how central BI teams should structure Microsoft Fabric workspaces, weighing medallion architecture patterns against deployment complexity, and debating whether centralized control helps or hinders self-service at scale—while unpacking why AI adoption intensifies workspace governance instead of simplifying it.

Get in touch:

We made it 500!

Get in touch:

Mike & Tommy explore whether being late to the data game is actually an advantage in the AI era, questioning if organizations building fresh Fabric implementations can leapfrog legacy technical debt—or if they're doomed to repeat the same governance mistakes. They tackle Sandra's mailbag question about late adoption, semantic modeling for AI readiness, and when to let ontology "brew" versus act now.

Get in touch:

Mike & Tommy tackle the classic BI trap of falling in love with tools before understanding the problem, exploring why organizations rush to adopt Copilot, Fabric, or dashboards without defining what decision they're trying to improve, and how to force problem clarity without becoming the blocker when stakeholders just want to "see the data."

Get in touch:

Mike & Tommy tackle a listener's question about connecting local VSCode to remote Fabric Jupyter kernels, exploring whether Fabric's architecture is fundamentally incompatible with traditional data science workflows, and discussing what an effective development cycle actually looks like when you can't just "pip install -e ." your way to productivity.

Mike & Tommy tackle a listener's challenge with Microsoft's overuse of "experience" in Fabric, questioning whether calling notebooks, pipelines, and copy jobs "experiences" inflates expectations and confuses adoption, and exploring what precise language would better serve data teams building real solutions.

Get in touch:

Mike & Tommy dive into Microsoft Fabric licensing strategy, questioning whether organizations actually need Fabric or are falling into the same over-provisioning trap as Power BI Premium. Drawing on Blake Edwards' client savings case studies, they explore F64 thresholds, hidden costs, and practical optimization tactics that separate smart capacity decisions from expensive assumptions.

Get in touch: